Alavipanah, Seyedkazem
Comparison between the correlation coefficients in salt crust area and crusted surface
reveals that when the desert crust areas are ignored from the analysis, the correlation
coefficients between TM bands 5 and 7 decreases. It means that TM bands 5 and 7 are
highly decorrelated in salt crust area. The result of FS analysis also showed that
dominant types of following salt crust can be recognized :
a- efflorescent, very fragile crystalline which appears to be snow covered ,
b- a salt - silt- clay crust is massive, relatively hard, whitish gray and usually has a
rough surface,
c- non- saline soil with a surface crust of a few millimeters thick and bright surface.
The results of PCA transformation on the crusted surface and salt crust are shown in
table 5 and 6 respectively. The eigenvectors computed for the correlation matrix on the
crusted surface indicates that TM visible bands 2,3,4 and TM band 7 have the most
contribution to the PC1 and TM bands 5 and 7 have the highest relationship with PC2.
No contribution of TM band 1 was found to the PC,S, that is due to the constant DN
values of 255 in the TM band 1 on the study area. The highest contribution to PC3 is
related to TM thermal band for both areas of salt crust ( table 6 ) and crusted surface
(table 5). Based on the obtained results, we may generally label PC1 as surface
appearance component, PC2 as moisture component. Table 5 shows that PC1 accounts
for 62.74% of the total variance and PC2 accounts for 31.94% of the variance.
Table 5. Eigenvectors of the seven components of the Qom crusted surface.
TM-1 |TM-2 JTM-3 |TM4 |TMS [TM.6 |TM-7 | Vom
PCI. 00 [0.504 [0.507 10.473 |-0.039 |-0.276 | 0.240 [62.74
PC? 0.0 10.10 0.086 |0.221 [0.564 105% |0.576 |31.94
PC3 00 10178 16175 (0419 [0.316 10790 (0.193 14.3
PC4 00 |-0.517 |-0.319 |0.068 |0.148 |-0:235 |-0.381 |0.51
Table 6. Eigenvectors of the seven components of the Qom salt crust.
TM- 1 TM-2 | TM-3 | TM-4. [TM-5:117M-6 | TM-7 | Yan
PC1 0.0 |0.483 |0.476 | 0.496 | 0.246 0.232 165.68
PC2 00 |-0.167 |-0.206 |0.005 | 0.641 0.649 | 30.09
PC3 00 10323 10271 10.260 [-0.089 10.850 |-0154 12.93
PC4 00 |-0246 10.193 |-0.428 |-0.545 10.036 [0.648 |0.86
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.
43